Ambiguous figures, fixation durations, and the bus paradox

Perceptual switching, or the reversal of ambiguous figures, is an extensively studied phenomenon in visual perception. Eye-movement studies have investigated the relation between switching behaviour and fixation duration, the interval between saccades. Ellis-Stark (1978) compared the average length of the interval in which a switch is reported to that of the surrounding intervals. They found that fixation durations are longer for intervals that contain a switch than for surrounding intervals. These observations, however, are highly susceptible to a sampling error called the 'bus paradox'. It did not report any bias control. A replication of their experiment is presented, in which bias control was applied to the fixation durations. The result shows an effect opposite to Ellis-Stark's approach; fixations in which a switch is reported are shorter than surrounding fixations.